A Practical Logging Pattern for Application Insights: Start, Flow, End, and Fail

When production issues happen, logs should clearly show what started, what happened in between, and how it ended. A simple and consistent pattern using _logger with Application Insights can make troubleshooting much easier.

Here is a practical example for an operation like retrieving role assignments.

Start – Log inputs clearly

Log the action and key parameters at the beginning:

const string action = "Retrieve role assignments for scope";
var stopwatch = Stopwatch.StartNew();

_logger.LogInformation("Start {Action}. Scope={Scope}, RoleId={RoleId}",
    action, scope, roleId);

This creates structured properties (Action, Scope, RoleId) that are searchable in Application Insights.

Flow – Log important checkpoints

Log meaningful steps inside the method:

_logger.LogInformation("Flow {Action} – calling external service. Scope={Scope}",
    action, scope);

Only log major steps. Avoid too many details.

End – Log result summary and duration

When successful:

stopwatch.Stop();

_logger.LogInformation("End {Action}. Count={Count}, ElapsedMs={ElapsedMs}",
    action, result.Count, stopwatch.ElapsedMilliseconds);

Include result summaries (like counts) and execution time. This helps detect performance issues.

Fail – Log exception with full context

On failure:

catch (Exception ex)
{
    stopwatch.Stop();

    _logger.LogError("Fail {Action}. Scope={Scope}, RoleId={RoleId}, ElapsedMs={ElapsedMs}, Exception={Exception}",
        action, scope, roleId, stopwatch.ElapsedMilliseconds, ex);

    throw;
}

Always pass the exception to LogError. Application Insights will capture the stack trace automatically.

Why This Works

• Messages are easy to read
• Properties are structured and searchable
• Start, Flow, End, and Fail follow the same pattern
• Operations teams can trace full execution paths

With this consistent approach, logs become a reliable story of what your system did — and why.